Speaker Identification, Differentiation and Verification Using Deep Learning for Human Machine Interface
Prakhar Mishra, Jabed Akhtar Choudhury, Edison Kho, Banani Basu, Arnab Nandi · 2023
Speaker distinction is critical in criminology and security organizations. Speaker differentiation is a form of voice recognition. The problem of precise differentiation is challenging. Researchers used various models to compare and classify voice comparison problems. But when it comes to accuracy deep learning is gaining traction once trained with the data in bulk for orator recognition and differentiation. This article presents a model that is efficient and accurate enough to segregate two or many voices. Also, the model finds the level of similarity of different voice samples. Even if the same speaker speaks twice after an age gap or with different emotions, the model can detect it as a different utterance of the same speaker. This paper exploits Deep learning- based methods for identification, fake speech detection, and orator diarization. The paper is giving insight into the past practices for speaker differentiation using deep learning. This short study will help novices and researchers comprehend the field of speech recognition and comparison.